Towards effective shell modelling with the FEniCS project

Size: px

Start display at page:

Download "Towards effective shell modelling with the FEniCS project"

Lee Grant
5 years ago
Views:

1 Towards effective shell modelling with the FEniCS project *, P. M. Baiz Department of Aeronautics 19th March 2013 Shells and FEniCS - FEniCS Workshop 13 1

2 Outline Introduction Shells: chart shear-membrane-bending and membrane-bending models example forms Two proposals for discussion: geometry: chart object for describing shell geometry discretisation: projection/reduction operators for implementation of generalised displacement methods Summary Shells and FEniCS - FEniCS Workshop 13 2

3 So far... dolfin manifold support already underway, merged into trunk [Marie Rognes, David Ham, Colin Cotter] I have already implemented locking-free (uncurved) beams and plate structures using dolfin manifold Next step: curved surfaces, generalised displacement methods (?) Aim of my talk is to start discussion on the best path Shells and FEniCS - FEniCS Workshop 13 3

4 Why shells? The mathematics: Shells are three-dimensional elastic bodies which occupy a thin region around a two-dimensional manifold situated in three-dimensional space The practical advantages: Shell structures can hold huge applied loads over large areas using a relatively small amount of material. Therefore they are used abundantly in almost all areas of mechanical, civil and aeronautical engineering. The computational advantages: A three-dimensional problem is reduced to a two-dimensional problem. Quantities of engineering relevance are computed directly. Shells and FEniCS - FEniCS Workshop 13 4

5 Figure : British Museum Great Court. Source: Wikimedia Commons. Shells and FEniCS - FEniCS Workshop 13 5

6 Figure : Specialized OSBB bottom bracket. Source: bikeradar.com Shells and FEniCS - FEniCS Workshop 13 6

7 A huge field There are many different ways of: obtaining shell models representing the geometry of surfaces on computers discretising shell models successfully And therefore we need suitable abstractions to ensure generality and extensibility of any shell modelling capabilities in FEniCS. Shells and FEniCS - FEniCS Workshop 13 7

8 Two methodologies Mathematical Model approach: 1. Derive a mathematical shell model. 2. Discretise that model using appropriate numerical method for description of geometry and fields. Degenerated Solid approach: 1. Begin with a general 3D variational formulation for the shell body. 2. Degenerate a solid 3D element by inferring appropriate FE interpolation at a number of discrete points. 3. No explicit mathematical shell model, one may be implied. Shells and FEniCS - FEniCS Workshop 13 8

9 Mathematical model Shells and FEniCS - FEniCS Workshop 13 9

10 Mathematical model shear-membrane-bending (smb) model Find U V smb : h 3 A b (U, V) + ha s (U, V) + ha m (U, V) = F(V) V V smb (1) membrane-bending (mb) model Find U V mb : h 3 A b (U, V) + ha m (U, V) = F(V) V V mb (2) Shells and FEniCS - FEniCS Workshop 13 10

11 Discretisation smb models vs mb models smb model takes into account the effects of shear; closer to the 3D solution for thick shells, matches the mb model for thin shells. Boundary conditions are better represented in smb model; hard and soft supports, boundary layers. smb U H 1 (Ω) vs mb U H 2 (Ω) Shells and FEniCS - FEniCS Workshop 13 11

12 Mathematical model Let s just take a look at the bending bilinear form A b for the mb model: A b (U, V) = ρ αβ H αβγδ b ρ γδ da (3a) Ω ρ αβ := ϕ,αβ t 1 j (u,1 (ϕ,2 t) u,2 (ϕ,1 t)) + 1 j (u,1 (ϕ,αβ ϕ,2 u,2 (ϕ,αβ ϕ,1 )) u,αβ t H αβγδ b := Eh 3 ( ) 12(1 ν 2 ν(ϕ,α ϕ,β ) +... ) (3b) (3c) Shells and FEniCS - FEniCS Workshop 13 12

13 Geometry Continuous model Terms describing the differential geometry of the shell mid-surface. The mid-surface is defined by the chart function. Discrete model We do not (usually) have an explicit representation of the chart. It must be constructed implicitly from the mesh and/or data from a CAD model. There are many different ways of doing this. Shells and FEniCS - FEniCS Workshop 13 13

14 Geometry Proposal 1 A base class Chart object which exposes various new symbols describing the geometry of the shell surface. Specific subclasses of Chart will implement a particular computational geometry procedure. The user can then express their mathematical shell model independently from the underlying geometrical procedure using the provided high-level symbols. Shells and FEniCS - FEniCS Workshop 13 14

15 shell_mesh = mesh (" shell. xml ") normals = MeshFunction (...) C = FunctionSpace ( shell_mesh, " CG", 2) chart = Chart ( shell_mesh, C, method =" patch_averaged ") chart = Chart ( shell_mesh, C, method =" CAD_normals ", normals = normals )... b_cnt = chart. contravariant_ basis () b_cov = chart. covariant_ basis () da = chart. measure () a = chart. first_ fundamental_ form ()... A_b =... Shells and FEniCS - FEniCS Workshop 13 15

16 Current discretisation options mb model: H 2 (Ω) conforming finite elements DG methods smb model: straight H 1 (Ω) conforming finite elements mixed finite elements (CG, DG) generalised displacement methods Shells and FEniCS - FEniCS Workshop 13 16

17 A quick note For simplicity I will just talk about the smb model reduced to plates, the chart function is the identity matrix; considerably simpler asymptotic behaviour but concepts apply to shells also. Shells and FEniCS - FEniCS Workshop 13 17

18 Locking 10 0 e L t = 0.1 t = 0.01 t = dofs Locking Inability of the basis functions to represent the limiting Kirchhoff mode. Shells and FEniCS - FEniCS Workshop 13 18

19 Move to a mixed formulation Treat the shear stress as an independent variational quantity: γ h = λ t 2 ( z 3h θ h ) S h Discrete Mixed Weak Form Find (z 3h, θ h, γ h ) (V 3h, R h, S h ) such that for all (y 3h, η, ψ) (V 3h, R h, S h ): a b (θ h ; η) + (γ h ; y 3 η) L 2 = f (y 3 ) t ( z 3h θ h ; ψ) L 2 2 λ (γ h; ψ) L 2 = 0 Shells and FEniCS - FEniCS Workshop 13 19

20 Move back to a displacement formulation Linear algebra level: Eliminate the shear stress unknowns a priori to solution [ ] { } { } A B u f B T = (5) C γ 0 To do this we can rearrange the second equation and then if and only if C is diagonal/block-diagonal we can invert cheaply giving a problem in original displacement unknowns: (A + BC 1 B T )u = f (6) Shells and FEniCS - FEniCS Workshop 13 20

21 Currently, this can be done with CBC.Block TRIA0220, TRIA1B20 [Arnold and Brezzi][Boffi and Lovadina] David Ham, Kent Andre-Mardal, Anders Logg, Joachim Haga and myself A, B, BT, C = [ assemble (a), assemble (b), assemble (bt), assemble (c)] K = collapse ( A - B * LumpedInvDiag ( C) * BT) Shells and FEniCS - FEniCS Workshop 13 21

22 Move back to a displacement formulation Variational form level: γ h = λ t 2 Π h ( z 3h θ h ) (7) Π h : V 3h R h S h (8) giving: a b (θ h ; η) + (Π h ( z 3 θ); y 3 η) L 2 = f (y 3 ) (9) Shells and FEniCS - FEniCS Workshop 13 22

23 Proposal 2 A new class Projection in UFL that signals to FFC that a projection between FunctionSpace objects is required. This requires additions in DOLFIN, UFL, FFC and FIAT. Shells and FEniCS - FEniCS Workshop 13 23

24 MITC7 Shells and FEniCS - FEniCS Workshop 13 24

25 ... V_3 = FunctionSpace (mesh, "CG", 2) R = VectorFunctionSpace ( mesh, " CG", 2, dim =2) + VectorFunctionSpace (mesh,"b", 3, dim =2) S = FunctionSpace ( mesh, " N1curl ", order =2) Pi_h = Projection ( from =R, to =S)... U = MixedFunctionSpace ([R, V_3 ]) theta, z_3 = TrialFunctions ( U) eta, y_3 = TestFunctions ( U) a_s = inner ( grad ( z_3 ) - Pi_h ( theta ), grad ( y_3 ) - Pi_h ( eta ))*dx Shells and FEniCS - FEniCS Workshop 13 25

26 Summary A big field with lots of approaches; need appropriate abstractions (inc. other PDEs on surfaces) Proposal 1: Expression of geometric terms in shell models using a natural form language which reflects the underlying mathematics Proposal 2: Effective discretisation options for the implementation of generalised displacement methods Shells and FEniCS - FEniCS Workshop 13 25

27 Thanks for listening. Shells and FEniCS - FEniCS Workshop 13 26

FEniCS Course. Lecture 8: A posteriori error estimates and adaptivity. Contributors André Massing Marie Rognes

FEniCS Course. Lecture 8: A posteriori error estimates and adaptivity. Contributors André Massing Marie Rognes FEniCS Course Lecture 8: A posteriori error estimates and adaptivity Contributors André Massing Marie Rognes 1 / 24 A priori estimates If u H k+1 (Ω) and V h = P k (T h ) then u u h Ch k u Ω,k+1 u u h